|
چکیده
|
The paper addresses the critical challenge of credit card fraud detection in an increasingly digital financial landscape. With the rise of online transactions, traditional fraud detection methods have proven insufficient against sophisticated fraudulent activities. This study proposes an innovative approach that integrates supervised learning and deep learning techniques, focusing on the use of autoencoders. Autoencoders, as unsupervised neural networks, are designed to learn efficient representations of transaction data, enabling the identification of anomalies indicative of fraud. By training these models on historical data, the research aims to enhance detection accuracy while minimizing false positives, thereby improving the overall effectiveness of fraud detection systems in real-time scenarios.
Furthermore, the research highlights the necessity for advanced methodologies in combating credit card fraud, especially given the alarming statistics indicating that nearly half of credit card users experienced fraud in 2022. The integration of autoencoders with supervised learning techniques not only addresses the challenges posed by imbalanced datasets but also leverages the strengths of both paradigms to create robust detection models. This hybrid approach is essential for adapting to evolving fraudulent tactics and maintaining consumer trust in digital payment systems. Ultimately, this study contributes to the ongoing efforts in enhancing fraud detection capabilities, ensuring financial institutions can better protect consumers from unauthorized transactions while fostering a secure transactional environment.
|